On Wed, April 30, 2008 19:28, karl shah-jenner wrote: > > ----- Original Message ----- > From: "David Dyer-Bennet" > > : Probably. There's a tradeoff, spinning wears the disk (rather slowly; > : look at the estimated lifespan for modern disks!), but starting the disk > : spinning is a lot of extra wear. So it depends how many starts vs. how > : many hours of spinning. And the exact numbers for any given drive > aren't > : really known and aren't available even as estimates. > : > : Powering a system down (well, it's more the powering it *up* step) has > an > : even bigger impact. Again, sitting vs. starting tradeoff, with the > : numbers not known. 20 years ago it was pretty clearly better to leave a > : system running for 24 hours rather than subject it to one extra power > : cycle. I'd expect that period of time to have been reduced since then, > : but haven't seen recent estimates. > : > > I didn't quote the whole article earlier when I posted this but I'll cite > one line again > "other notable patterns showed that failure rates are indeed definitely > correlated to drive manufacturer, model, and age; failure rates did not > correspond to drive usage except in very young and old drives " > > > it's a recent study by one of the biggest single user of drives. Right, I should have thought of the Google drive survey, I've heard about it and seen quotes before. In fact what I *should* do is go read it, since I'm interested in the topic and interested in actual facts. Sounds like the physical tech has advanced to the point that the behavior is significantly different -- I'm sure they haven't magically caused running *not* to wear the bearings, or starts *not* to stress the motors and bearings, but it sounds like those items have gotten so good that except in special cases they're not main causes of failure any more. -- David Dyer-Bennet, dd-b@xxxxxxxx; http://dd-b.net/ Snapshots: http://dd-b.net/dd-b/SnapshotAlbum/data/ Photos: http://dd-b.net/photography/gallery/ Dragaera: http://dragaera.info